A biosensor is a device that uses biological recognition properties for the selective analysis of various analytes or biomolecules. Generally, the sensor will produce a signal that is quantitatively related to the concentration of the analyte. In particular, a great deal of research has been directed toward the development of a glucose sensor that would function in vivo to monitor a patient's blood glucose level. Such a glucose sensor is useful in the treatment of diabetes mellitus. In particular, an implantable glucose sensor that would continuously monitor the patient's blood glucose level would provide a physician with more accurate information in order to develop optimal therapy. One type of glucose sensor is the amperometric electrochemical glucose sensor. Typically, an electrochemical glucose sensor employs the use of a glucose oxidase enzyme to catalyze the reaction between glucose and oxygen and subsequently generate an electrical signal. The reaction catalyzed by glucose oxidase yields gluconic acid and hydrogen peroxide as shown in the reaction below (equation 1):
  glucose  +            O      2        ⁢          →      oxidase      glucose        ⁢    gluconic    ⁢                  ⁢    acid    +            H      2        ⁢          O      2      The hydrogen peroxide reacts electrochemically as shown below in equation 2:H2O2→2H++O2+2e−
The current measured by the sensor is generated by the oxidation of the hydrogen peroxide at a platinum working electrode. According to equation 1, if there is excess oxygen for equation 1, then the hydrogen peroxide is stoichiometrically related to the amount of glucose that reacts with the enzyme. In this instance, the ultimate current is also proportional to the amount of glucose that reacts with the enzyme. However, if there is insufficient oxygen for all of the glucose to react with the enzyme, then the current will be proportional to the oxygen concentration, not the glucose concentration. For the glucose sensor to be useful, glucose must be the limiting reagent, i.e., the oxygen concentration must be in excess for all potential glucose concentrations. Unfortunately, this requirement is not easily achieved. For example, in the subcutaneous tissue the concentration of oxygen is much less that of glucose. As a consequence, oxygen can become a limiting reactant, giving rise to a problem with oxygen deficit. Attempts have been made to circumvent this problem in order to allow the sensor to continuously operate in an environment with an excess of oxygen.
Several attempts have been made to use membranes of various types in an effort to design a membrane that regulates the transport of oxygen and glucose to the sensing elements of glucose oxidase-based glucose sensors. One approach has been to develop homogenous membranes having hydrophilic domains dispersed substantially throughout a hydrophobic matrix to circumvent the oxygen deficit problem, where glucose diffusion is facilitated by the hydrophilic segments.
For example, U.S. Pat. No. 5,322,063 to Allen et al. teaches that various compositions of hydrophilic polyurethanes can be used in order to control the ratios of the diffusion coefficients of oxygen to glucose in an implantable glucose sensor. In particular, various polyurethane compositions were synthesized that were capable of absorbing from 10 to 50% of their dry weight of water. The polyurethanes were rendered hydrophilic by incorporating polyethyleneoxide as their soft segment diols. One disadvantage of this invention is that the primary backbone structure of the polyurethane is sufficiently different so that more than one casting solvent may be required to fabricate the membranes. This reduces the ease with which the membranes may be manufactured and may further reduce the reproducibility of the membrane. Furthermore, neither the percent of the polyethyleneoxide soft segment nor the percent water pickup of the polyurethanes disclosed by Allen directly correlate to the oxygen to glucose permeability ratios. Therefore, one skilled in the art cannot simply change the polymer composition and be able to predict the oxygen to glucose permeability ratios. As a result, a large number of polymers would need to be synthesized before a desired specific oxygen to glucose permeability ratio could be obtained.
U.S. Pat. Nos. 5,777,060 and 5,882,494, each to Van Antwerp, also disclose homogeneous membranes having hydrophilic domains dispersed throughout a hydrophobic matrix to reduce the amount of glucose diffusion to the working electrode of a biosensor. For example, U.S. Pat. No. 5,882,494 to Van Antwerp discloses a membrane including the reaction products of a diisocyanate, a hydrophilic diol or diamine, and a silicone material. In addition, U.S. Pat. No. 5,777,060 to Van Antwerp discloses polymeric membranes that can be prepared from (a) a diisocyanate, (b) a hydrophilic polymer, (c) a siloxane polymer having functional groups at the chain termini, and optionally (d) a chain extender. Polymerization of these membranes typically requires heating of the reaction mixture for periods of time from 1 to 4 hours, depending on whether polymerization of the reactants is carried out in bulk or in a solvent system. Therefore, it would be beneficial to provide a method of preparing a homogenous membrane from commercial polymers. Moreover, as mentioned above, one skilled in the art cannot simply change the polymer composition and be able to predict the oxygen to glucose permeability ratios. Therefore, a large number of polymers would need to be synthesized and coating or casting techniques optimized before a desired specific oxygen to glucose permeability ratio could be obtained.
A further membrane is disclosed in U.S. Pat. No. 6,200,772 B1 to Vadgama et al. that has hydrophilic domains dispersed substantially throughout a hydrophobic matrix for limiting the amount of glucose diffusing to a working electrode. In particular, the patent describes a sensor device that includes a membrane comprised of modified polyurethane that is substantially non-porous and incorporates a non-ionic surfactant as a modifier. The non-ionic surfactant is disclosed as preferably including a poly-oxyalkylene chain, such as one derived from multiple units of poly-oxyethylene groups. As described, the non-ionic surfactant may be incorporated into the polyurethane by admixture or through compounding to distribute it throughout the polyurethane. The non-ionic surfactant is, according to the specification, preferably incorporated into the polyurethane by allowing it to react chemically with the polyurethane so that it becomes chemically bound into its molecular structure. Like most reactive polymer resins, complete reaction of the surfactant into the polyurethane may never occur. Therefore, a disadvantage of this membrane is that it can leach the surfactant over time and cause irritation at the implant site or change its permeability to glucose.
PCT Application WO 92/13271 discloses an implantable fluid measuring device for determining the presence and the amounts of substances in a biological fluid that includes a membrane for limiting the amount of a substance that passes therethrough. In particular, this application discloses a membrane including a blend of two substantially similar polyurethane urea copolymers, one having a glucose permeability that is somewhat higher than preferred and the other having a glucose permeability that is somewhat lower than preferred.
An important factor in obtaining a useful implantable sensor for detection of glucose or other analytes is the need for optimization of materials and methods in order to obtain predictable in vitro and in vivo function. The ability of the sensor to function in a predictable and reliable manner in vitro is dependent on consistent fabrication techniques. Repeatability of fabrication has been a problem associated with prior art membranes that attempt to regulate the transport of analytes to the sensing elements.
We refer now to FIG. 1, which shows a photomicrograph at 200× magnification of a prior art cast polymer blend following hydration. A disadvantage of the prior art membranes is that, upon thermodynamic separation from the hydrophobic portions, the hydrophilic components form undesirable structures that appear circular 1 and elliptical 2 when viewed with a light microscope when the membrane 3 is hydrated, but not when it is dry. These hydrated structures can be detected by photomicroscopy under magnifications in the range of between 200×-400×, for example. They have been shown by the present inventors to be non-uniform in their dimensions throughout the membrane, with some being of the same size and same order of dimensions as the electrode size. It is believed that these large domains present a problem in that they result in a locally high concentration of either hydrophobic or hydrophilic material in association with the electrode. This can result in glucose diffusion being limited or variable across the dimension adjacent the sensing electrode. Moreover, these large hydrated structures can severely limit the number of glucose diffusion paths available. It is noted that particles 4 in membrane 3 are dust particles.
With reference now to a schematic representation of a known membrane 14 in FIG. 2A, one can consider by way of example a continuous path 16 by which glucose may traverse along the hydrophilic segments 10 that are dispersed in hydrophobic sections 12 of the membrane. For path 16, glucose is able to traverse a fairly continuous path along assembled hydrophilic segments 10 from the side 18 of the membrane in contact with the body fluid containing glucose to the sensing side 20 proximal to sensor 22, where an electrode 24 is placed at position 26 where glucose diffusion occurs adjacent surface 20. In particular, in that portion of the membrane 14 proximal to position 26, glucose diffusion occurs along hydrophilic segments 10 that comprise a hydrated structure 28 having a size and overall dimensions x that are of the same order of magnitude as electrode 24. Therefore, glucose diffusion would be substantially constant across the dimension adjacent electrode 24, but the number of glucose diffusion paths would be limited.
Referring now to FIG. 2B, one can consider an example where glucose traversing prior art membrane 14 from side 18 in contact with the body fluid to the sensing side 20 cannot adequately reach electrode 30. In particular, electrode 30 is located at position 34, which is adjacent to a locally high concentration of a hydrophobic region 12 of prior art membrane 14. In this instance, glucose diffusion cannot adequately occur, or is severely limited across the dimension adjacent the electrode surface. Consequently, one would expect that the locally high concentration of the hydrophobic regions adjacent to working electrode 30 would limit the ability of the sensing device to obtain accurate glucose measurements. The random chance that the membrane could be placed in the 2A configuration as opposed to 2B leads to wide variability in sensor performance.
We also refer to FIG. 2C, which shows another cross-section of prior art membrane 14. In this instance, glucose is able to traverse a fairly continuous path 36 from side 18 to side 20 proximal to the sensing device. However, electrode 38 is located at position 40 such that glucose diffusion is variable across the dimension adjacent the electrical surface. In particular, most of the electrode surface is associated with a locally high concentration of hydrophobic region and a small portion is associated with hydrophilic segments 10 along glucose diffusion path 36. Furthermore, glucose diffusing along path 36a would not be associated with the electrode. Again, the large non-uniform structures of the prior art membranes can limit the number of glucose diffusion paths and the ability of the sensing device to obtain accurate glucose measurements.
It would be beneficial to form more homogeneous membranes for controlling glucose transport from commercially available polymers that have a similar backbone structure. This would result in a more reproducible membrane. In particular, it is desired that one would be able to predict the resulting glucose permeability of the resulting membrane by simply varying the polymer composition. In this way, the glucose diffusion characteristics of the membrane could be modified, without greatly changing the manufacturing parameters for the membrane. In particular, there is a need for homogeneous membranes having hydrophilic segments dispersed throughout a hydrophobic matrix that are easy to fabricate reproducibly from readily available reagents. Of particular importance would be the development of membranes where the hydrophilic portions were distributed evenly throughout the membrane, and where their size and dimensions were on an order considerably less than the size and dimensions of the electrode of the sensing device to allow the electrode to be in association with a useful amount of both hydrophobic and hydrophilic portions. The ability of the membranes to be synthesized and manufactured in reasonable quantities and at reasonable prices would be a further advantage.